Much of our work at SMCSE helps organizations adopt digital transformation, data and AI tools. The Cloud Has a Body: What AI’s Climate Story Leaves Out looks at what sits underneath those tools: the data centers, power grids and water supply that AI runs on.
The report was published by Effective Thoughts, a knowledge platform for research-driven content across economics, technology, and society. SMCSE Solutions and Wisdom Inn for Skill Enhancement (WISE) collaborated on it.
What organizations using AI should know
- Every AI query has a physical footprint. Behind it sit server racks, cooling systems, transmission lines, and the mines that supply the minerals inside every chip. Global data center electricity demand is on track to more than double by 2030, reaching roughly 945 terawatt-hours.
- Most of the water use happens off-site. Roughly 60% of a typical data center’s water footprint comes from generating the electricity it draws, not from on-site cooling.
- Reporting rules are uneven. In the EU, data centers above 500kW must report energy and water data, yet only 36% did. In the US, the federal government has been removing environmental review while states add their own rules.
- Capacity moves across borders. When Singapore paused new data center approvals between 2019 and 2022, investment moved to Johor, where operational capacity roughly doubled from about 401 MW in 2023 to around 897 MW in 2024. Malaysia’s water regulator found that of over 800 million liters a day requested by proposed projects, only about 142 million can realistically be supplied.
- Better is possible. In Denmark, Meta’s Odense data center recovers about 215,000 MWh of waste heat a year, warming more than 12,000 homes through the local district heating network.
Read the full report
The complete report, The Cloud Has a Body: What AI’s Climate Story Leaves Out, is available on Effective Thoughts.